Analysis Of Classification And Tracking In Vehicles Using Shape Based Features

Vehicle classification plays a prominent role in Electronic Toll Collection (ETC). The reason behind any method to incorporate into the system of ETC is to reduce the time at toll plazas and even the safety. Even many methods had been derived and implemented still we find some tackling and classification related problems. In this paper, a new algorithm has been proposed to avoid tackling and classification problems. In the proposed algorithm first the frame difference method is used to detect the moving regions from the highway scene. Some morphological operations are used to remove the shadow noise and to detect the moving object correctly.  After vehicle  detection, a  region-based  vehicle  tracking method  is  used for  building  the  correspondence  between  vehicles  detected  at different  time  instants. After vehicle tracking, we consider two parameters such as aspect ratio and compactness are used to classify and count the vehicles. Experimental results on different videos with different lighting conditions are used to demonstrate the effectiveness of our proposed system.